The Message Design Logics of Responses to HIV Disclosures
John P. Caughlin, Dale E. Brashers and Mary E. Ramey
University of Illinois at Urbana-Champaign
Kami A. Kosenko
North Carolina State University
Erin Donovan-Kicken
University of Texas at Austin
Jennifer J. Bute
Ohio University
[This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x]
Author Note
The order of the last four authors is reverse alphabetical.
Corresponding author: John Caughlin, email: [email protected].
Responses to HIV Disclosures 2
Abstract
This manuscript uses the theory of message design logics to investigate the relative sophistication of
responses to disclosure of HIV status. In Study 1, 548 college students imagined a sibling revealing an
HIV-positive diagnosis. Their responses to the HIV-disclosures were coded as expressive (n = 174),
conventional (n = 298), or rhetorical (n = 66). Type of message produced was associated with gender and
HIV aversion. In Study 2, 459 individuals living with HIV rated response messages that were taken
verbatim from Study 1. Expressive messages were rated lowest in quality, and rhetorical messages were
rated highest. The discussion focuses on the utility of message design logics for understanding responses
to HIV disclosures and the implications for message design logics.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 3
The Message Design Logics of Responses to HIV Disclosures
Issues surrounding the disclosure of an HIV-positive status have become prominent concerns in
the health communication literature (for a review, see Greene, Derlega, Yep, & Petronio, 2003). People
often fail to disclose that they are living with HIV because of the stigma associated with the illness and
concerns about how others will react (e.g., Alonzo & Reynolds, 1995; Herek, Capitanio, & Widaman,
2002), but keeping the diagnosis a secret can lead to a lack of social support, unsafe sexual behavior, and
failure to seek treatment or to take medications as needed (e.g., Chesney & Smith, 1999). Stigma and lack
of disclosure of HIV status are such important health concerns that they warrant counseling programs to
help these individuals manage the psychological, social, and health aspects of their illness (Rintamaki,
Davis, Skripkauskas, Bennett, & Wolf, 2006).
Much HIV disclosure research has examined how people living with the disease decide whether
to divulge their status (e.g., Derlega, Winstead, & Folk-Barron, 2000; Derlega, Winstead, Greene,
Serovich, & Elwood, 2004; Sullivan, 2005). Some research also has examined disclosers’ perceptions of
the responses they receive when telling others about their HIV-positive diagnosis (Greene & Faulkner,
2002; Serovich, Kimberly, & Greene, 1998). Greene and Faulkner, for example, interviewed 10 female
African American adolescents, who described how people responded when told about their HIV
diagnosis. The responses varied considerably and included treating the participants unfavorably, having
negative emotional reactions, telling others, and providing support.
Understanding how people respond to disclosures about HIV-positive status is important for
numerous reasons. For instance, people living with HIV often experience uncertainty about how others
will react (Brashers et al., 2003). If this uncertainty contributes to anxiety about revealing the information,
research describing reactions to HIV disclosures can be useful by illuminating “the issues and dilemmas
that may be encountered when disclosing” (Serovich et al., 1998, p. 15).
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 4
Research on responses to HIV disclosures from the perspective of the responder is also important.
People who disclose an HIV-positive test result may hope for various sorts of responses from the
recipient, including offers of social support (Derlega, Lovejoy, & Winstead, 1998; Greene & Faulkner,
2002), assurances that others will not be told (Greene & Faulkner, 2002), and a chance for catharsis
(Derlega et al., 1998). Presumably, at least some recipients of HIV disclosures want to provide high
quality responses, but it is not entirely clear what counts as “high quality” in this context. The current
literature provides some general guidance about competent responses, such as not dismissing the HIV
disclosure as unimportant (Barbee, Derlega, Sherburne, & Grimshaw, 1998). Still, there has been little
systematic attention on how to conceptualize the competence or quality of responses to HIV disclosures.
Thus, the primary objective of the present research is to proffer a theoretical means of
conceptualizing message sophistication in responses to disclosures of HIV-positive test results. Our
discussion is rooted in O’Keefe’s (1988, 1990, 1991) theory of message design logics. The second
objective is to extend research on message design logics, illustrating how this research can be applied to
novel research contexts. The majority of research on message design logics to date has focused on
interpersonal influence situations, but the current research demonstrates that the theory can be applied
usefully to responses to important disclosures – in this instance, the disclosure of an HIV-positive test
result. We pursued these objectives in the two studies described below, with the first focusing on
developing the model for classifying message sophistication and the second testing the model.
Message Design Logics Perspective
Numerous scholars have noted that communicators often encounter situations in which they have
multiple goals or aims (e.g., Clark & Delia, 1979; Dillard, 1990; Goldsmith, 2004; Wilson, 2002). The
obvious commonality among various multiple goals perspectives is the recognition that communicators
typically have multiple objectives when interacting. For instance, even when communicators’ most overt
aim is to persuade, their messages also shape how they are viewed and have implications for their
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 5
relationship with the other person (Clark & Delia). The theory of message design logics is a particular
multiple goals perspective that is associated with a specific set of theoretical propositions. First, the
message design logic theory posits that various messages can be viewed as means of addressing
communicative goals. In this view, goals are “the central elements in socially codified representations of
situations” (O’Keefe, 1988, p. 82). The “design logic” is what connects goals to messages (i.e., messages
are designed in ways that support the logic of how to accomplish various goals).
O’Keefe (1988) conceptualized goals as being implicit in social structures. That is, certain goals
normatively are accepted as relevant to particular situations; there is a conventional understanding that
such objectives should be pursued and met. Brown and Levinson (1987), for instance, argued that the
concerns about the other person’s identity are an inherent risk of making requests. In particular, asking
someone to comply with a request inevitably threatens that person’s negative face, which refers to the aim
of remaining unfettered by impositions on one’s autonomy. Different goals are relevant to other
situations: Refusing another individual’s request can interfere with one’s identity goals, such as seeming
cooperative -- or at least not difficult (Saeki & O’Keefe, 1994); people run the risk of appearing nosey if
they give advice (Goldsmith & Fitch, 1997; Kunkel, Wilson, Olufowote, & Robson, 2003); and offers of
social support can leave the recipient feeling a loss of control over his or her circumstances (Brashers,
Neidig, & Goldsmith, 2004).
Because goals are conceptualized in terms of their relevance to particular situations, the relevance
of goals is not dependent on whether specific individuals identify with those goals (O’Keefe, 1988, 1992).
Indeed, “individuals sometimes do fail to adopt and pursue goals that are intrinsically relevant to a
situation” (O’Keefe, 1988, p. 82), but such cases can illustrate the conventional relevance of the goals. If
a person does not attend to the risk of appearing nosey while giving advice, for example, the target of the
advice may take offense and evaluate the advice-giver negatively (Goldsmith & Fitch, 1997). It is
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 6
normatively expected that advice-givers will attempt to meet the identity goal of not being overly
intrusive, and this goal remains pertinent whether or not any single communicator meets it.
The most distinctive feature of the message design logics theory is that it posits that there are at
least three separate means by which objectives can be pursued in communicative situations. These
different means are not simply strategies; instead, they reflect three distinct sets of premises regarding
how one attempts to meet communicative goals (O’Keefe, 1988). The most elementary design logic is
called expressive. The basic principle of the expressive design logic is “language is a medium for
expressing thoughts and feelings” (O’Keefe, 1988, p. 84). Messages reflecting an expressive logic focus
on openness, honesty, and clear self-expression. Expressive messages also tend to focus on responding to
the preceding events rather than orienting toward meeting intrinsic objectives of the situation; that is, by
focusing on expressing prominent thoughts and feelings, expressive messages may fail to attend to certain
normatively salient communicative goals. Expressive messages also may contain insults, unconditional
threats (i.e., ones without contingencies for the recipient to avoid the threatened behavior), substantial
redundancies, and other inappropriate comments (O’Keefe, 1988).
The next design logic, conventional, is rooted in the premise that “communication is a game
played cooperatively, according to socially conventional rules and procedures” (O’Keefe, 1988, p. 86).
Like the expressive view, conventional messages involve expressing propositions, but unlike the
expressive view, messages with conventional design logic do not always involve expressing whatever
thoughts and feelings are prominent to the speaker. Conventional messages instead prioritize doing what
is appropriate within the current social circumstances; thus, speakers utilizing a conventional design logic
attend to the normatively relevant goals, and they may use various politeness strategies to mitigate any
potential inappropriateness. Conventional messages treat social structures, such as role positions and
obligations, as fixed. In short, conventional messages attempt to meet the normative goals in a given
social situation, and they attempt to do so in conventionally appropriate ways.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 7
The third design logic, rhetorical, has the premise that “communication is the creation and
negotiation of social selves and situations” (O’Keefe, 1988, p. 87). The rhetorical design logic includes
the social knowledge inherent in the conventional logic, but it does not treat social situations as fixed.
Instead, the rhetorical logic presumes that the context is constituted through communication and that
communicators can renegotiate the social situation. Messages using the rhetorical design logic utilize
various features of communication that suggest certain identities and relationships for the interactants.
Producers of rhetorical messages seek to foster a desirable interpersonal consensus about the social
situation. Because such negotiations require consensus to be successful, rhetorical messages seek
harmony; thus, they do not include conflict resolution techniques like enforcing power relations or
blaming the other. Instead, “rhetorical message producers tend to join the partner’s project or counter with
their own rather than simply giving a conventionally appropriate response” (O’Keefe, 1988, p. 88).
The three design logics are ordered developmentally with expressive being the least sophisticated
and rhetorical being the most (O’Keefe, 1988). This ordering is based on theoretical grounds: The ability
to utilize an expressive logic is required before one can master conventional usage, and thorough mastery
of conventional communication is needed before one can successfully renegotiate identities and situations
(O’Keefe, 1988). There also has been ample evidence that the same order generally corresponds to
competence. O’Keefe and McCornack (1987), for instance, asked undergraduates to evaluate persuasive
messages and found that rhetorical ones were rated highest in effectiveness and competence, and
expressive ones were rated the lowest. Similar results have been found in health communication contexts:
Lambert and Gillespie (1994) found that hypertension patients who received messages about complying
with blood pressure medication regimens rated rhetorical messages as most effective and attractive.
Message design logics also have been linked to other known indicators of communicative competence;
for example, Peters (2005) found that sophistication of design logic in comforting messages was
associated with the message producer’s cognitive complexity. In short, there are compelling theoretical
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 8
and empirical grounds for considering the design logics to be ordered in level of sophistication, from
expressive (least) to rhetorical (most).
Suggesting that rhetorical messages are the most sophisticated does not imply that they are
always necessary or desirable. The general superiority of rhetorical messages over conventional ones (and
conventional ones over expressive ones) is limited to “complex communicative situations” (O’Keefe,
1988, p. 91). Such situations involve multiple pertinent goals, including competing goals that are difficult
to achieve simultaneously. There are many fairly simple communicative situations (e.g., normative
conversations with the grocery store cashier) in which one would not expect great variety in the design of
messages. In comparatively simple situations, the advantage of using more sophisticated strategies would
not be realized because the situational objectives may not conflict. One would not expect there to be a
particular advantage to sophisticated design logics, for example, if the scenario were calling a friend just
to “kill some time” (see Hullman, 2004).
To date, the theory of message design logics usually has been applied to situations in which the
message producer attempts to regulate the other person’s behavior (e.g., Lambert & Gillespie, 1994;
O’Keefe, 1988, 1990). Regardless of whether the behavior involves taking medication (Lambert &
Gillespie) or getting a group member to do a fair share of work (O’Keefe, 1988), such regulative
situations are inherently complex because asking for another person’s compliance is a threat to his or her
autonomy (Brown & Levinson, 1987). Peters (2005) recently applied message design logic theory to
scenarios in which women and their spouses coped with breast cancer, but even this research involved a
number of scenarios involving attempts to influence the other's behavior (such as the wife asking the
husband to do more housework). However, because Peters also included scenarios in which comforting
was clearly warranted (e.g., some husbands were asked how they would respond if their wife was feeling
unattractive due to cancer treatments), his research demonstrates that message design logics can be
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 9
applied usefully beyond regulative situations, to a broader range of scenarios of interest to communication
researchers. One such communicative situation is being told of an HIV-positive test result by a sibling.
Relevant Goals in HIV-Disclosure Situations
Because the message design logics vary in the extent to which (and means by which) they attend
to normatively relevant goals, applying message design logics to a new social circumstance should begin
with a consideration of the goals that are relevant to that type of scenario (see O’Keefe & Shepherd,
1987). The literature on HIV disclosures provides a detailed account of the types of aims and concerns
people have when considering whether to reveal their HIV-positive status (e.g., Derlega et al., 2000;
Derlega et al., 2004; Greene et al., 2003; Serovich, Mason, Bautista, & Toviessi, 2006; Sullivan, 2005).
One prominent reason that people consider withholding information about an HIV-positive status
is they are worried about being thought of negatively, because of the stigma often associated with the
disease. Individuals considering disclosing their HIV-positive status may expect negative emotional
reactions (Greene & Faulkner, 2002) or even feel ashamed themselves (Derlega et al., 1998; Derlega et
al., 2004). In the vernacular of multiple goals theories, people deciding whether to disclose HIV-positive
test results often are concerned about maintaining a positive identity. This implies that people who do
reveal their HIV-positive status would have the goal of minimizing any loss of positive identity.
Relational goals are also important to people considering whether to disclose HIV-positive test
results. In one study, the most frequently cited reason for not disclosing was fear of rejection, including
concerns that the target would no longer like the discloser (Derlega et al., 1998). Thus, maintaining a
positive relationship with the disclosure recipient (or at least not worsening the existing relationship) is
likely a salient goal when revealing an HIV-positive status. Persons living with HIV also sometimes balk
at telling others due to privacy issues (Derlega et al., 2000; Derlega et al., 2004). If one is concerned that
“people have big mouths and they might go running around telling other people” (Derlega et al., 2000, p.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 10
68), he or she may hesitate divulging information about HIV status. And, if one does disclose, he or she
may be interested in minimizing the chances of subsequent disclosures.
There are also some common factors that impel people toward disclosure. Based on the research
by Derlega and his colleagues (2004), the two most salient reasons for disclosing are probably feeling a
duty to inform the other (e.g., not wanting the person to be surprised by the information later) and having
a close and supportive relationship. Disclosing because one has a supportive relationship implies that
people often disclose an HIV-positive test result because they expect that “the other person would be
understanding, consoling, concerned, and supportive” (Derlega et al., 1998, p. 153). That is, receiving
emotional support is a common expectation of those who would disclose, and it is reasonable to infer that
soliciting support is a common goal for people disclosing.
The general goals relevant to disclosing HIV, which are summarized above, need to be
understood within a particular social context. The reasons for revealing an HIV-positive status differ
depending on the relationship with the target. For instance, divulging an HIV-positive status due to
loyalty appears to be more prominent with family members than with relational partners, but a desire for
honesty and a need to protect the other’s health are cited more frequently with revelations to relational
partners than to family members (Derlega et al., 1998).
There are reasons to believe that HIV disclosures to family members may have complexities and
risks not evident in disclosures to friends or even romantic partners. Family relationships typically are
nonvoluntary, and often are viewed as particularly difficult to replace (Vangelisti, 1993). Also, people are
more likely to regret revealing their HIV-positive status to family members than they are to regret
revealing to friends (Serovich et al., 2006). Such findings indicate that disclosing to family members is a
complex and potentially risky situation that warrants investigation. In the present study, we focused on
disclosures to a particular family member, siblings. Our decision to ask participants about siblings was
based on the need to have participants think about a relationship that would be salient to most of them.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 11
Also, although disclosing to siblings is often a particularly important challenge for people living with
HIV, disclosures in that relationship are not well understood (Greene et al., 2003). Moreover, because
siblings are in the same generation, the power discrepancies tend to be less than those of some other
family relationships, like parent-child relationships. Avoiding large power discrepancies between the
participant and discloser was important because power differences could create overly strict constraint on
how participants might respond. People might feel compelled to withhold negative evaluations of a
parent, for instance, but sibling relationships are likely less subject to such response constraints.
Our discussion of goals to this point has focused on the types of goals that are likely relevant to
people disclosing HIV-positive test results. This reflects the research literature, which has focused more
on the person disclosing than recipients of such disclosures. Nevertheless, it is possible to infer likely
goals for people whose sibling divulges an HIV-positive test result to them. Recipients of a sibling’s HIV
disclosure may have a variety of idiosyncratic goals in that situation. Some, for instance, may wish to
avoid their sibling due to feeling anxious about HIV (e.g., Le Poire, 1994), but the nonvoluntary nature of
the sibling relationship may make such avoidance difficult or undesired. Moreover, research examining
the conversational dynamics of painful self-disclosures suggests that such conversations necessarily entail
complex and problematic responsibilities for recipients, including whether and how to enable or inhibit
additional disclosive talk (Coupland, Coupland, & Giles, 1991; Coupland, Henwood, Coupland, & Giles,
1989). Coupland and colleagues (1991, 1989) analyzed the variety of conversational "next moves" a
recipient might make in response to a personal and sensitive revelation and concluded that recipients
themselves face competing and potentially contradictory goals, such as responding in a sympathetic and
appropriate manner without escalating talk about an uncomfortable topic (e.g., how can recipients express
sympathy, yet simultaneously make it clear that they do not want to discuss this topic anymore?).
Regardless of their personal goals, however, there are also some conventionally relevant goals in
a scenario in which one’s sibling reveals an HIV-positive test result. Such information certainly is
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 12
considered “bad news,” and having one reveal such information typically constitutes a scenario in which
providing emotional support is expected. That is, having a sibling divulge that he or she has HIV is a
situation that would be normatively recognized as one calling for comforting. As the literature on
comforting makes clear, this does not mean that every individual will recognize the situational objective
of comforting and act in a suitably supportive manner (see, e.g., Burleson, 1994). Indeed, the failure to
provide emotional support in situations that are conventionally understood to require it partly defines
what constitutes poor or incompetent emotional support. One necessary characteristic of a conventionally
competent response to having a sibling disclose an HIV-positive test result, therefore, would be some act
that would normatively be understood as comforting.
Study 1
The primary objective of the first study was to provide a theoretically derived means for
conceptualizing the competence and sophistication of responses to HIV disclosures. There were
theoretical reasons to expect that O’Keefe’s (1988) theory of message design logics could serve as a
useful way of classifying the sophistication of responses. Yet, because the situation in the present study is
so distinct from the bulk of research using message design logics, it was important to ask two very basic
research questions:
RQ1: Will the responses to siblings' HIV disclosures vary in sophistication as defined by the use
of message design logics?
RQ2: If the responses do vary in sophistication, can those variations be described in a manner to
allow for reliable ratings of the variations?
Examining whether the response messages varied in design logics and whether those variations
could be reliably rated was the first step in the first study. The second step involved making predictions
about how the messages with different design logics would be related to other constructs. First, previous
research on message design logics has shown that women overall produce messages with more
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 13
sophisticated design logics than do men (O’Keefe, 1988). There is also evidence that, on average, women
are more skillful at providing emotional support than are men (e.g., MacGeorge, Gillihan, Samter, &
Clark, 2003). Thus, we expected
H1: Compared to men, women will construct messages with more sophisticated design logics.
Also, one of the defining characteristics of HIV-disclosure situations is the stigma often
associated with the virus (Alonzo & Reynolds, 1995; Herek et al., 2002). Recipients of HIV disclosures
certainly vary in the extent to which they endorse that stigma, with some finding the disease so distasteful
that they are averse to interacting with people who have HIV and others not sharing such concerns (e.g.,
Le Poire, 1994). Some people find HIV so aversive that they even fear touching someone who is HIV-
positive. Such individuals likely would experience fear for their own health if a sibling would disclose an
HIV-positive test result. Whereas some individuals who experience HIV aversion may be able to
overcome those emotions and produce comforting responses to HIV disclosures, severe aversion to HIV
may make people focus on their own emotional state, likely increasing the chances that they would
produce expressive messages and decreasing the chances that they would produce anything more
sophisticated than a conventionally appropriate response. Being averse to HIV also may inhibit one’s
motivation to provide an appropriate and effective comforting response. In short, there were several
theoretical reasons to expect that
H2: The extent of HIV aversion will be related inversely to the sophistication of the messages
constructed.
The message design logics perspective also suggests that variations in HIV disclosure messages
may influence how recipients of those messages respond. O'Keefe (1988) argued that some messages can
redefine the social situation, including the identities of those involved. Thus, different disclosure
messages may create or constitute slightly different situations that may elicit varying responses.
Obviously, the number of possible HIV disclosure messages is potentially infinite, but the aforementioned
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 14
research on HIV disclosures (e.g., Agne, Thompson, & Cusella, 2000; Derlega et al., 2004; Schrimshaw
& Siegel, 2002) suggests a number of common goals to such disclosures, including (a) disclosing the
information, (b) seeking social support, (c) avoiding negative evaluation or stigma, (d) preventing
subsequent disclosure to a third party, (e) maintaining the relational bond with the other person, and (f)
telling because the other person had a right to know the information. Attending (or not attending) to such
goals could change the tenor of an HIV disclosure, potentially influencing the sophistication of the
response. Consequently, we posed the following research question:
RQ3: Do HIV disclosure messages that vary in their attention to particular goals elicit responses
that differ in sophistication?
Method
Participants. We enlisted 581 participants from communication courses at a large Midwestern
university to complete questionnaires. Due to the focus of the investigation (i.e., sibling disclosure of
HIV status), we asked interested participants if they had at least one sibling before distributing
questionnaires. If participants had more than one sibling, they were instructed to complete the
questionnaire with their oldest sibling in mind. Thirty-three participants reported having no siblings, and
those individuals completed a separate questionnaire about a cousin. Because the sample reporting on
cousins was not large enough for systematic examination, the questionnaires concerning cousins were
excluded from the analysis, leaving a sample size of 548.
Of the final sample, 299 (54.6%) were female, 239 (43.6%) were male, and 10 (1.8%) provided
no information on their gender. Participants averaged 19.72 years of age (SD = 1.85, minimum = 18,
maximum = 39). The majority of the sample was non-Hispanic White (n = 392, 71.5%), with the
remainder reporting as African American (n = 56, 10.2%), Asian American (n = 40, 7.3%), and Hispanic
(n = 30, 5.5%). Seventeen participants (3.1%) reported other ethnic origins.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 15
Procedures. Each participant was given a sheet of paper with a single, open-ended question and
an envelope containing a longer questionnaire with a variety of closed-ended items. The open-ended
question read, “Imagine you are alone talking with your sibling. (If you have more than one sibling, pick
your oldest sibling.) After a few minutes of small-talk, your sibling says, ‘Well, I have something to tell
you…’” This statement was followed by 1 of 24 disclosure messages indicating that the sibling had HIV.
After the disclosure statement, participants were asked, “If this were to occur, what would you do or say
in response? If you would say something, write down exactly the words you would use.” Participants
wrote their responses on the single sheet of paper containing the open-ended question and, when finished,
put the sheet of paper in the envelope provided. Then they completed a series of closed-ended questions
relevant to the HIV-disclosure scenario. The primary purpose of the envelope was to prevent participants’
responses on the open-ended question from being influenced by the closed-ended items in the
questionnaire. The measures relevant to the present manuscript are described below.
Questionnaires and coding. Each questionnaire began with a hypothetical situation in which a
sibling revealed an HIV-positive test result to the participant. The 24 specific disclosure messages were
grouped into six conditions, which reflected the common disclosure goals summarized above. The four
messages, which focused only on disclosing the information, served as a kernel condition (Jackson,
1992); that is, in addition to forming their own conditions, these messages were embedded as part of the
larger messages for the other five conditions. In addition to including a kernel message, the messages in
each of the remaining conditions explicitly referred to one additional goal. For the second through sixth
conditions, respectively, these goals were seeking support, avoiding negative evaluation, preventing
subsequent disclosure, maintaining the relationship with the recipient, and honoring the recipient's right to
know the information. More details about these conditions, including a list of the exact messages, can be
found in Caughlin et al. (in press).
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 16
The assessment of message design logic was based on the open-ended responses to the
disclosures. Of the 548 participants, 539 completed the open-ended response to the disclosure. The
response messages were examined in a holistic manner to determine whether there were meaningful
variations in design logic. All six of the authors read subsamples of the responses and agreed that the
messages could be described usefully using O’Keefe’s (1988) message design logics.
Prior to coding, all of the authors collaborated on adapting the message design logic categories to
the current research problem; for instance, based on the argument that the goal of providing social support
is inherently relevant when a sibling reveals a positive HIV status, messages that did not include at least a
minimal attempt at social support were said to fail at engaging an important goal that is inherently
relevant to the situation. Such failures are one of the diagnostic hallmarks of expressive messages
(O’Keefe, 1988).
After the adaptation phase, four of the authors coded each of the messages into one of three
categories: expressive (i.e., messages that treat communication as medium for conveying thoughts and
feelings, often without attending to a conventionally relevant situational goal), conventional (i.e.,
messages that treat communication as a rule-based system that is followed in a normative manner), and
rhetorical (i.e., messages treat communication as an activity that constitutes the social situation, often
attempting to define or redefine identities or relationships). Reliability among the coders was assessed
with an intraclass correlation (Fleiss, 1981), and it was excellent: .91. Discrepancies in the coding were
resolved in meetings with at least five of the authors present. The coding rules can be found in the
Appendix.
The closed-ended questionnaires included assessments of demographic information. Additionally,
HIV aversion was assessed with a series of 7-point Likert-type items asking participants how they
generally would react to people with HIV. The four items for this measure were taken from Le Poire’s
(1994) study of nonverbal stigmatization of people with HIV/AIDS, in which the measure demonstrated
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 17
adequate reliability and validity. This HIV tolerance measure taps the participants’ fearful and
stigmatizing reactions toward individuals with HIV/AIDS. Sample items include: “I would not eat with a
person who has HIV” and “I would touch someone who has HIV.” Positively phrased items were
recoded, so that higher scores on the four items indicated a tendency toward HIV aversion. After
examining the scale reliability (α = .75), all four items were retained (M = 2.15, SD = 1.39).
Also, a subsample (n = 325) of the participants were asked to rate the realism of the disclosure
message. Two items tapped the realism of the disclosure scenario (e.g., “Think about how your sibling
disclosed in the hypothetical situation described above. To what extent was this similar to how somebody
with HIV might disclose this to a sibling?”), and each ranged from 1 to 7 with higher scores being more
realistic (α = .77). On average, participants reported that the messages were fairly realistic, with a mean of
4.65 (SD = 1.38), which was significantly greater than the scale midpoint, t (324) = 8.50, p < .001.
Results
In overview, our results demonstrate that responses to illness disclosure can be categorized as
predicted by the theory of message design logics. That is, the answers to the first two research questions
were both affirmative: The responses varied in the sophistication of the message design logics (RQ1), and
it was possible to adapt the message design logic theory to reliably classify the variations in design logics
(RQ2). As shown in Table 1, conventional messages were most common, followed by expressive and
then rhetorical responses (differences between the categories of design logics were all statistically
significant using the binomial test, p < .001). The following sections briefly overview the main themes of
the different types of messages and report results of our hypothesis tests.
Although our analysis of message design logics was based on the whole response messages, the
examples presented below are subdivided into thought units (Hatfield & Weider-Hatfield, 1978). The
divisions provide a means for commenting on specific parts of the messages. Also, the quotation marks
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 18
that appear with some examples indicate where participants wrote quotations marks as part of their
response.
Expressive design logic. Expressive messages revealed the emotions and thoughts of the
respondent, but typically lacked comforting or offers of social support (which were more common in
conventional responses -- see the following section). These messages revealed the disclosure recipient’s
emotional or cognitive state, such as sadness, disbelief, or need to elicit more information about the
disclosure, as shown in Examples 1 and 2.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 19
Example 1
01: “Oh my God! 02: Are you serious? 03: I cannot believe this. 04: How did this happen? 05: Are you sure? 06: You’ve been to the doctor and everything?!”
Example 2
01: What? 02: How’d you get it? 03: How long have you known? 04: So what’s this mean? 05: Well…that fuckin’ sucks.
Other expressive messages represented explicitly negative responses, as in Example 3, which
reflects the respondent’s anger at the discloser (line 1), and includes blame (lines 2 and 3) and
stigmatizing remarks about contagion (line 4). Example 4 has similar features, but also includes an
additional hurtful comment about the harm that disclosure will do to others (line 4).
Example 3
01: You are fucking kidding me! 02: What the hell did you do? 03: If it was your own fault, then I have no pity. 04: Be careful when you are around me.
Example 4
01: Are you fucking kidding me? 02: Who have you told yet? 03: Well you better go tell Mom and Dad. 04: You know you are going to break mom’s heart. 05: When did you get tested? 06: Well you should get tested again to make sure.
If expressive messages did have an attempt at comforting or support, they had some element that
negated it, such as explicit blame. Example 5 below contains minimal comforting and an offer of
assistance (line 5), but that follows blame for the infection (lines 2 and 3) as well as questions about the
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 20
relationship (line 4). This message also reveals the respondent’s own emotional state, another expressive
act (line 1).
Example 5
01: First of all, I would probably start crying. 02: Then I would ask her, “how did this happen? 03: I didn’t even know you were participating in such risky behavior! 04: I thought we told each other everything!” 05: But I would make sure she knew I love her and would do anything to help her.
Conventional design logic. Conventional responses represent a baseline for normative behavior.
For example, the typical response for a self-disclosure of illness would be comforting and offering
support (i.e., these would be “intrinsically relevant” goals, see O’Keefe, 1988). Comforting or support
messages may be more or less sophisticated (see Goldsmith, 2004), but conventional messages should, as
a minimum, contain an attempt to console, calm, reassure, or offer assistance to the discloser. Example 6
explicitly addresses comforting as a goal of the message (line 1) and an attempt to offer help (line 3) and
elicit the emotional state of the discloser (line 4).
Example 6
01: I would ask my brother questions to better understand his condition and to help comfort if he is upset. 02: I would first ask, “Have you told Ann?” (his wife) 03: I would also ask, “In what ways can I help you most in this time?” 04: and “How are you doing emotionally with this news?”
In both example 6 and example 7 (below), the offers of support are general (e.g., what can I do to
help?), rather than specific plans (which are more indicative of rhetorical responses – see the following
section).
Example 7
01: Oh my gosh! 02: When did you find this out? 03: Is there anything I can do to help? 04: I’m very sorry that this happened to you.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 21
Rhetorical design logic. Rhetorical messages typically were marked by attention to multiple
goals. Example 8 shows efforts to comfort (lines 1 and 2), followed by contextualizing the situation (lines
4 though 7).
Example 8
01: My initial response would be to say “I’m so sorry” and just hug my sister. 02: We are not a very physical family, so this act of physical support would be important. 03: Then, I would say, “Well, this is really a shock, 04: but life will go on. 05: Focus as much as you can about making the best out of this situation. 06: You’ve been such a positive person all your life. 07: I have confidence that you will make this a positive situation as well. 08: Just have faith that God will help you and guide you.” 09: Past that, I would definitely ask a lot of questions, 10: and figure out medically and emotionally how she is doing.
Rhetorical responses sometimes also included signifiers of a “communal coping” orientation (Lyons et
al., 1999), including the use of “we” statements (e.g., line 8 “we’ll get through this…”). Such statements
suggest an attempt to redefine the circumstances as a joint project (see O’Keefe, 1988).
Example 9
01: I support you, John. 02: You can count on me. 03: I’ll encourage you to take your medications, 04: doctor’s appointments you need to go to. 05: If your physical condition requires it, I’ll take care of your chores. 06: I just want you to concentrate on following your medical treatment to strengthen your health first. 07: Remember, we are a family 08: & we’ll get through this because I care & love you.
In summary, the three design logics were represented in our data, and we were able to reliably
code messages into these categories. Expressive messages simply reflected the thoughts and feelings of
the respondent (e.g., convey sadness, shock, anger), conventional messages attended to socially normative
behavior (e.g., offer comforting and support), and rhetorical messages attended to context and planning
(e.g., understand context and make specific plans for coping with the problem). This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 22
Hypotheses and third research question. The first hypothesis predicted that women overall would
produce messages with more sophisticated design logics than would men. The relevant frequencies for
this hypothesis are summarized in Table 1. An initial chi-square test suggested that there likely was some
association between messages with different design logics and the sex of the message producer, χ2 (2) =
4.98, p = .08, φ = .10. To determine if any differences were consistent with the hypothesis, we conducted
tests for significant differences between the proportions of each message logic for men versus women.
Such tests of differences in proportions are distributed as z (see Brunig & Kintz, 1997), and the specific
percentages are listed in Table 1. Overall, men were more likely than women to produce expressive
messages, z = 1.99, p < .05. Women were more likely than men to produce conventional messages, z = -
2.17, p < .05. Both of these significant findings are consistent with the hypothesis; however, there was not
a significant difference between men and women with respect to rhetorical messages, z = 0.46, ns.
The second hypothesis involved the expectation that participants’ level of HIV aversion would be
related inversely with the sophistication of their message. This hypothesis was examined with a one-way
analysis of variance, comparing people who produced messages with different design logics on HIV-
aversion, F (2, 322) = 2.43, p = .09, η2 = .01. Consistent with the hypothesis, post hoc analyses indicated
that people who wrote expressive messages (M = 2.37, SD = 1.49) were higher than people with
conventional messages (M = 2.00, SD = 1.26) in terms of HIV-aversion, p < .05, η2 = .02. Those with
expressive messages were also higher in HIV-aversion than people who fashioned rhetorical messages (M
= 2.11, SD = 1.54), but this difference was not significant.
To determine whether the six disclosure message conditions influenced the sophistication of
responses (RQ3), we examined the message design logics of the responses for each disclosure condition
(see Table 2). Because the message design logics are ordered in terms of sophistication, a series of Mann-
Whitney U tests was conducted. Generally, the condition asking to prevent subsequent disclosure to a
third party received the least sophisticated responses, and this condition was significantly lower than the
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 23
kernel (U = 3015, z = 2.89, p < .01), seeking-support (U = 3191, z = 1.97, p < .05), avoiding-negative-
evaluation (U = 3241, z = 2.00, p < .05), and maintaining-relationship (U = 3041, z = 2.24, p < .05)
conditions.
These overall differences between the prevent-subsequent-disclosure condition and other
conditions were explored further by comparing the proportions of the three design logics for each of the
conditions. Messages involving preventing subsequent disclosure elicited more expressive messages than
did any other type of message, and the proportion of expressive design logics was significantly higher for
the prevent-subsequent-disclosure condition than for the kernel (z = 2.56, p < .05) and the maintain-
relationship conditions (z = 2.16, p < .05). There was also evidence that prevent-subsequent-disclosure
messages elicited expressive replies more than did seeking-support messages (z = 1.95, p = .05). The
prevent-subsequent-disclosure condition was about average in terms of conventional messages and was
not significantly different from any other condition in conventional responses. Prevent-subsequent-
disclosure evoked the fewest rhetorical messages, with two conditions, the kernel (z = 1.97, p < .05) and
messages about honoring the right to know (z = 2.04, p < .05), significantly higher. Also, there was some
evidence that the condition concerning avoiding negative evaluation elicited more rhetorical responses
than did the messages about preventing subsequent disclosure (z = 1.91, p = .06).
As summarized in Table 2, there were three additional significant differences in the proportions
of design logics. Specifically, the kernel messages had fewer expressive responses than did the right-to-
know messages (z = 2.12, p < .05), and the right-to-know condition had fewer conventional responses
than did the seeking-support (z = 2.22, p < .05) or maintain-relationship (z = 2.23, p < .05) conditions.
Because these findings do not reflect any overall differences in the aforementioned Mann-Whitney U-
tests, they should be interpreted with caution. Indeed, although the right-to-know condition was fairly
high in expressive responses, it also elicited rhetorical replies as frequently as any other condition (and
more frequently than the prevent-subsequent-disclosure condition).
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 24
Discussion of Study 1
The current study provides an important conceptual advance in our understanding of what counts
as a sophisticated response to an HIV disclosure. This research demonstrates that it is possible to reliably
classify response messages using O’Keefe’s (1988) message design logics. Our adaptation of the message
design logics theory provides a useful description of what competent and incompetent messages look like.
Because this classification system is rooted in a hierarchical ordering that has been confirmed in several
studies (e.g., O’Keefe & McCornack, 1987; Peters, 2005), it provides a coherent way to classify and
assess the sophistication of responses.
The utility of the current adaptation of O’Keefe’s (1988) model is supported by the findings
pertaining to the hypotheses in the present investigation. Consistent with previous research on message
design logics (O’Keefe, 1988), we found that men were more likely to produce relatively unsophisticated
messages than were women. Replicating the finding that women are more prone to produce relatively
sophisticated messages provides evidence of the validity of our adaptation: Given that past research has
shown such systematic sex differences, valid new assessments of message sophistication would be
expected to generate similar differences as well.
It also should be noted that O'Keefe's (1988) findings indicated that women produced more
rhetorical messages than did men, whereas in our study, men produced more expressive messages than
did women and women produced more conventional messages than did men. One should be cautious
about interpreting such discrepancies between studies too finely because some of the apparent differences
are based on comparing a significant result in one study to a null finding in the other. It is unclear whether
the null results really indicate no meaningful sex differences; for instance, O'Keefe's study included only
20 expressive messages, and even though men and women produced expressive messages in "roughly
equal numbers" (p. 95) the proportion of men producing expressive messages was 28% compared to only
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 25
18% of the women. Nevertheless, the differences in the studies with respect to sex suggest that the
specific task in the current study is somewhat different from that used by O'Keefe.
Such differences are not surprising. O’Keefe (1988) argued that different types of tasks may
require more or less sophisticated design logics (e.g., simple requests like asking someone to pass the salt
may require less attention to face needs). In addition to the slightly different results pertaining to sex, our
distribution of design logics was significantly different than O’Keefe’s (1988), χ2 (2) = 24.78, p < .001, φ
= .04. Her data yielded 22% expressive, 46% conventional, and 33% rhetorical messages (compared to
32%, 55%, and 13% respectively in our data). It is likely that differences in the type of task (i.e.,
regulative situation versus response to an HIV disclosure) or the nature of the task (e.g., explicitness of
goals) could account for such differences. For example, it may be that the disclosure of an HIV-positive
diagnosis calls so strongly for comforting, that conventional messages are more common. On the other
hand, people may have such a strong negative reaction (e.g., anger or anxiety) that expressive messages
also are more common. Designing a rhetorical message in this situation may be the most challenging,
because of the simultaneous needs such as managing stigmatized identities, protecting vulnerable family
relationships, supporting a sibling with a life-threatening illness, and processing bad news.
Also, the findings pertaining to the varying sophistication of responses to different disclosure
messages (RQ3) differed from previous work on message design logics. Previous work on message
design logics has tended to highlight the individual differences in message design that are evident in
particular complex situations (e.g., Lambert & Gillespie, 1994; O’Keefe, 1988; O’Keefe, Lambert, &
Lambert, 1997). Message design logics theory, however, explicitly states that situations can influence the
messages produced; for instance, in a routine context, everyone may produce routine messages (O’Keefe,
1988). In the current study, responses to the prevent-subsequent-disclosure messages tended to evoke
comparatively frequent expressive responses and comparatively infrequent rhetorical ones. This suggests
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 26
that explicitly asking the recipient to withhold subsequent disclosures helps constitute a context that is not
conducive to eliciting responses with sophisticated design logics (see Table 2).
It is not clear from these data why the prevent-subsequent-disclosure condition would evoke
relatively unsophisticated responses. Other analyses from the same data, however, provide some clues
about this finding (see Caughlin et al., in press). Participants who responded to prevent-subsequent
disclosure messages were particularly likely to make statements blaming the sibling for acquiring HIV.
This suggests that asking the recipients not to divulge the information may heighten judgments of
responsibility and may make participants particularly cognizant of the stigma associated with HIV.
Drawing attention to perceived responsibility or stigma may diminish respondents' willingness or capacity
to make a well-developed response.
Also, we found that people responding to the prevent-subsequent-disclosure condition were less
likely than people responding to other conditions to report that they would feel closer to their sibling as a
result of such a disclosure (Caughlin et al., in press). Being told explicitly to not reveal the diagnosis
could threaten the recipient's identity because it implies that the recipient is untrustworthy. It also implies
a lack of trust in the relationship. Recipients may believe that their trustworthiness as a confidant should
be unquestioned, and explicit instructions not to divulge the information may imply that the discloser
does not take the recipient's trustworthiness for granted. Such a threat to their identity could make
recipients less motivated to respond in a sophisticated manner. There clearly are other explanations, but
regardless of the particular reason why prevent-subsequent-disclosure messages evoke unsophisticated
responses, these results are important because they draw attention to the situational component of
message design logics, which has not often been highlighted in previous research.
The most obvious limitation to the first study is that it cannot provide direct evidence that the
hierarchical ordering of the message categories reflects the relative quality of the categories. Based on the
theory of message design logics (O’Keefe, 1988), expressive messages are least sophisticated (and
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 27
rhetorical messages are most sophisticated) in terms of addressing the multiple goals of the situation;
however, the first study cannot demonstrate that messages of greater sophistication are more effective or
appropriate for the particular context of responding to a sibling's HIV disclosure. This is an important
limitation because previous research has provided only mixed and indirect evidence for the notion that
sophisticated design logics enhance message quality in comforting situations (Peterson & Albrecht,
1996).
To address this limitation in the first study, the goal of the second study was to examine whether
the message categories differed in terms of perceived quality or competence. Competence often is
conceptualized as having effectiveness and appropriateness components (e.g., Canary & Spitzberg, 1987).
Given the normative salience of comforting to a scenario in which a sibling discloses an HIV-positive test
result, perceived effectiveness likely is based on the extent to which a response is viewed as comforting
(i.e., helpful, supportive, and sensitive; Goldsmith, McDermott, & Alexander, 2000). The hypotheses for
the second study were:
H3a: Expressive messages will be rated as lower in comforting and appropriateness than both
conventional messages and rhetorical messages.
H3b: Conventional messages will be rated as lower in comforting and appropriateness than
rhetorical messages.
Method
Participants. Because they would be uniquely positioned to judge the quality of responses to HIV
disclosures, we recruited participants who are HIV-positive. Specifically, we posted information about the
study on HIV-related Internet bulletin boards, and we mailed flyers to many of the HIV/AIDS service
organizations listed on www.thebody.com. A total of 459 individuals completed the study. Participants
who were willing to provide contact information were entered in a lottery to win 1 of 10 $75 prizes.
Participants resided in all regions of the US, including 41 states plus Washington DC. The average age in
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 28
the sample was 43.15 (SD = 9.22, minimum = 19, maximum = 74). Most of the participants reported
being non-Hispanic White (n = 244, 53.2%), African American (n = 150, 32.7%), or Hispanic (n = 43,
9.4%). The ethnic backgrounds of the remaining individuals varied considerably, with the largest single
group being Native Americans (n = 14, 3.1%). There were 325 men (70.8%), 122 women (26.6%), and 12
people (2.6%) who did not report gender or listed another descriptor of gender. Almost all (n = 453,
98.7%) of the respondents had at least one sibling. On average, participants reported first testing positive
for HIV 12.30 years ago (SD = 6.85, minimum = 1, maximum = 28). About half (n = 229, 49.9%) of the
respondents had been diagnosed with AIDS, and the average time since diagnosis was 9.81 years (SD =
5.38, minimum = 1, maximum = 25).
Procedures. Potential participants were directed to a webpage that contained information about
the study, including the pertinent information regarding informed consent. An Internet study was selected
because findings from Internet research are generally consistent with more traditional methods, but web-
based studies typically attract more diverse samples (Gosling, Vazire, Srivastava, & John, 2004). Those
choosing to participate selected a link to a secure survey site (www.surveymonkey.com). The link
between the consent page and the survey website was designed to direct participants randomly to one of
thirty different versions of the survey. To prevent large discrepancies in the number of people completing
the various versions of the survey, we monitored the number of people completing each version and
closed versions after 20 participants had completed them. At least 13 participants completed each version
of the questionnaire.
Questionnaires. The initial questions asked participants how long ago they first tested positive for
HIV and whether they had been diagnosed with AIDS. The participants then were asked to imagine that
they had just told a sibling that they were HIV-positive for the first time and to "imagine that your brother
or sister reacts in the way described below." The reactions that participants were asked to imagine were
taken from responses to the kernel messages in Study 1. We focused on responses to kernel messages
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 29
because many of the responses to the other disclosure messages included explicit references to the
disclosure content; for example, one person who was asked to ensure privacy wrote, "Of course I will
keep it quiet." Each respondent was asked to imagine one particular reaction out of a total of 30. The 30
reaction messages were selected by randomly picking 10 expressive responses, 10 conventional
responses, and 10 rhetorical responses. Other than the specific reaction message, the 30 versions of the
questionnaire were identical.
After being asked to consider the reaction message, participants rated the quality of the messages
using 16 semantic differential items with 5-point scales. The first 12 items were taken from Goldsmith et
al.'s (2000) comforting measure, which has three subscales: helpfulness, supportiveness, and sensitivity.
The reliabilities in the current sample were .89, .92, and .94, respectively. The remaining four items
("appropriate- inappropriate," "rude–decent," "respectful–disrepectful," and "proper-improper") assessed
appropriateness, and Cronbach's alpha was .92.
The respondents were also asked to rate the realism of the response message using a semantic
differential scale from unrealistic (1) to realistic (5). The average realism rating was 3.49 (SD = 1.01).
This was significantly greater than the scale midpoint, t (455) = 10.30, p < .001.
Results
Although participants generally found the scenarios to be realistic, it seemed prudent to make
sure that the primary results of the study could not be attributed to variations in perceived realism. Before
examining the hypotheses, we tested whether perceived realism was influenced by the message design
logics. Specifically, we conducted an ANOVA that treated the specific messages as a random factor
within the three messages design logics. This design accounts for variations that may occur because of
differences in messages within a particular category, and it makes it possible to generalize the findings
beyond the specific response messages examined to the larger categories of message design logics
(Jackson, 1992; Jackson & Brashers, 1994). Overall, there was not significant evidence that the message
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 30
design logics influenced the ratings of realism, F (2, 27) = 1.74, p = .19. There were, however, some
small associations between realism and the measures of message quality. The correlations between
realism and ratings of helpfulness, supportiveness, sensitivity, and appropriateness, respectively, were .12
(p = .01), .12 (p = .01), .09 (p = .06), and .13 (p < .01). Although these correlations represent small effect
sizes, it is conceivable that such associations could influence the primary analyses of the study; thus,
perceived realism was controlled in all subsequent analyses.
The tests of the hypotheses also involved treating the response messages as a random factor
within the three design logic categories. Additionally, perceived realism was added as a covariate, and we
included planned contrasts to test whether conventional messages were rated as more comforting and
appropriate than were expressive messages (H3a) and whether rhetorical messages were rated as more
comforting and appropriate than were conventional messages (H3b). The first comforting measure was
the helpfulness subscale, and there was an overall significant difference among the three message design
logics, F (2, 27) = 22.85, p < .001, partial- η2 = .63; η2 = .22. Because a covariate was included in the
model, the planned contrasts were based on estimated marginal means (EMM). Consistent with H3a, the
planned contrasts indicated that messages with an expressive logic (EMM = 2.89, SE = .07) were rated
lower in helpfulness than were messages with a conventional logic (EMM = 3.82, SE = .07, p < .001).
Although rhetorical messages (EMM = 3.95, SE = .07) were rated higher in helpfulness than conventional
messages, this difference was not statistically significant (p = .18), meaning H3b was not confirmed.
There was also an overall significant effect for design logics on ratings of supportiveness, F (2,
27) = 32.64, p < .001, partial- η2 = .71; η2 = .24. The planned contrasts were consistent with both parts of
the third hypothesis. Expressive messages (EMM = 2.92, SE = .07) were rated significantly lower in
supportiveness than were conventional messages (EMM = 3.89, SE = .07, p < .001), and conventional
messages, in turn, were significantly lower than rhetorical ones (EMM = 4.09, SE = .07, p = .05).
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 31
The results pertaining to sensitivity were also consistent with both parts of the third hypothesis.
The overall effect of the message design logics was significant, F (2, 27) = 23.28, p < .001, partial- η2 =
.63; η2 = .19. The planned contrasts indicated that expressive messages (EMM = 3.02, SE = .07) were
lower in sensitivity than were conventional messages (EMM = 3.83, SE = .07, p < .001) and that
conventional messages were rated as less sensitive than were rhetorical messages (EMM = 4.07, SE = .07,
p = .02).
Finally, there was a significant overall effect for message design logic on ratings of
appropriateness, F (2, 27) = 24.47, p < .001, partial- η2 = .64; η2 = .20. The planned contrasts showed that
expressive messages (EMM = 3.02, SE = .07) were lower in appropriateness than were conventional
messages (EMM = 3.87, SE = .07, p < .001). There was not, however, a significant difference between
conventional messages and rhetorical ones (EMM = 4.02, SE = .07, p = .11).
Discussion
The primary goal of Study 2 was to test the model from Study 1 that was developed to rank the
sophistication of messages responding to a sibling's disclosure of an HIV-positive test result. Although
the proposed hierarchy of messages design logics from Study 1 was rooted in O'Keefe's (1988) theory of
message design logics, Study 1 could not provide direct evidence that messages of greater sophistication
would be viewed as more comforting and appropriate. The results from Study 2, however, generally
confirmed that more sophisticated messages are rated as higher in quality. In particular, the participants
rated conventional messages as significantly more helpful, supportive, sensitive, and appropriate than
expressive messages. Additionally, rhetorical messages were rated higher than conventional ones on all
four measures, and these differences were statistically significant for supportiveness and sensitivity. In
short, the results from Study 2 confirm that the theory of message design logics provides a useful means
of conceptualizing and categorizing the quality of responses to a sibling's disclosure of a positive HIV test
result.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 32
General Discussion
It is important to understand what constitutes effective and appropriate responses to disclosures of
HIV-positive test results. Disclosing a positive HIV status can be important for seeking support,
managing health, and numerous other reasons. Yet there are often significant barriers to disclosure (e.g.,
Derlega et al., 2004). Poor (unsophisticated) responses may stifle disclosure even more: When a person
gets a bad response, he or she may choose not to reveal certain details about the diagnosis (e.g., specific
needs for assistance). Moreover, poor responses from one person can lead people living with HIV to
withdraw and be less likely to tell others. Patients even report stigmatizing responses from health care
providers, which diminishes their trust and confidence in the health care system (Brashers, Hsieh, Neidig,
& Reynolds, 2006).
The research presented here provides an important advance in our understanding of what
constitutes a high quality response to the disclosure of an HIV-positive test result. Consistent with
O'Keefe's (1988) theory of message design logics, messages that show greater sophistication in how they
attend to the relevant goals generally were rated as higher in quality by participants in Study 2. These
findings confirm that differences in message design logics are a useful way of understanding what makes
response messages more or less effective and appropriate in general.
This is not to suggest that the more sophisticated messages will always be evaluated most
favorably; it is axiomatic that messages are interpreted and evaluated in context. Thus, many factors may
moderate the overall tendency for more sophisticated messages to be viewed as higher quality. For
instance, O'Keefe et al.'s (1997) research suggests that individuals who strongly value open expression
may disapprove of message features associated with a rhetorical logic, such as attempts to renegotiate
social identities. Clearly, future research should examine such potential moderators as they could provide
caveats to our findings, which focus on providing a framework for understanding which type of messages
are viewed most favorably in general.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 33
In addition to providing a general framework for understanding what constitutes better responses
to HIV disclosures, the message design logic perspective highlighted a number of specific message
elements that can be translated into concrete advice about how to respond to HIV disclosures. Developing
coding guidelines for the response messages (see the Appendix) involved specifying particular message
features that were characteristic of the three design logics. These specific features imply specific advice
for constructing effective response messages. For example, given that expressive messages were rated
lowest in quality, individuals wishing to respond appropriately to someone's HIV disclosure would do
well to avoid message features that are invariably associated with an expressive design logic, such as
noncontingent threats, insults, explicit blame, or explicit criticism.
Although the primary purpose of the research reported here was to offer a theoretically grounded
procedure for understanding the relative sophistication of responses to HIV disclosures, the current
studies also have implications for our understanding of message design logics. Most past research on
message design logics has focused on interpersonal influence situations such as gaining compliance
(O'Keefe, 1990) or workplace conflicts (O'Keefe et al., 1997). The current investigations demonstrate that
message design logics can be applied to a very different social situation: responding to a sibling's
disclosure of an HIV-positive test result. The applicability of the message design logics framework to
such a markedly different context implies that the framework has much broader utility than might be
suggested by the relatively narrow applications heretofore.
There were also some theoretically interesting differences between the current research and
previous work on message design logics. Findings from the first study, for example, demonstrated that the
nature of the disclosure message can influence the sophistication of the response. Given that the second
study confirmed that message sophistication is related positively to perceived comforting and
appropriateness, this has important applied implications. People disclosing potentially stigmatizing
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 34
information, such as an HIV-positive test result, should be aware that the message used to divulge that
information can influence the other person's response.
The current study demonstrates that the sophistication of responses is influenced by features of
disclosure messages. Moreover, in other analyses, we found that different disclosure messages have
different effects on other aspects of participants' responses, such as the extent to which they would
evaluate the sibling negatively and the extent that they would feel closer to their sibling because of the
disclosure (Caughlin et al., in press). The findings are complex in that no single disclosure message type
is the most effective across all dimensions, but consistent with the current results, prevent-subsequent
disclosure messages were rated low on a number of dimensions (as were messages that honored the
recipient's right to know the information). Taken together, these results strongly suggest that the quality of
response messages is diminished by disclosures that explicitly request that there be no subsequent
disclosure.
Additionally, the findings presented here suggest that there were some important differences
between HIV disclosure situations and the situations examined in most previous research on message
design logics. In the first study, the proportion of rhetorical messages was smaller than that found in
previous work by O'Keefe (1988). In the second study, although the findings were generally consistent
with the hierarchy proposed by O'Keefe (1988), the differences in quality ratings between expressive
messages and conventional messages were more pronounced than those between conventional messages
and rhetorical messages (e.g., conventional messages were rated as superior to expressive ones on all four
indicators of comforting and appropriateness whereas rhetorical messages were rated as significantly
superior to conventional ones on only two indicators). Taken together, these findings suggest that the
situation of responding to an HIV disclosure is different from the previously examined interpersonal
influence situations. Given that the current situation elicits fewer rhetorical messages than the influence
situation and that rhetorical messages appear to confer only a slight advantage in quality over
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 35
conventional messages, it is plausible that what the current situation really demands is at least a
conventional response. That is, the normative situational goal of providing at least minimal comfort
(while avoiding the negative features that automatically make a response expressive) may be the
overriding imperative in the current situation.
Like any study, this one has limitations. First, it should be noted that receiving HIV disclosures
from a sibling is a specific situation that may not generalize to HIV disclosures in other types of
relationships. Second, the disclosures in the current study represented single episodes, but it is likely that
many sibling dyads would continue to discuss an HIV diagnosis over time. Consequently, although we
would expect that the initial reaction would remain important, understanding siblings' later responses may
also be important (e.g., even siblings who respond badly at first may become supportive later). Another
limit of the study is that we used hypothetical reactions rather than actual reactions to HIV-disclosures.
Care should be taken in generalizing these findings; however, there are good reasons to believe that
findings based on hypothetical scenarios can provide a useful theoretical basis for further understanding.
We were struck by the fact that the data in Study 1 included considerable variability in responses, many
commonalities within categories (i.e., participants seemed to have a stock of knowledge on which to
draw), and that the messages generated seemed quite realistic. Indeed, participants in Study 2, all of
whom are living with HIV, rated the response messages produced in Study 1 as generally realistic. Thus,
whereas the nature of the data is a limitation, there are ample grounds for considering the findings
meaningful for helping understand what constitutes a sophisticated response to an HIV disclosure.
In conclusion, disclosing one’s HIV-positive status can be challenging, and may be inhibited by
fears about how the target of the disclosure might react. This pair of studies demonstrates that a wide
range of responses are possible – from (a) relatively unsophisticated expressive messages that might
include blame, criticism, or the expression of the sadness or surprise of the disclosure target, to (b)
messages that address conventional norms, such as offering support and comforting, to (c) rhetorical
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 36
messages that define the situation, explicitly address identity and relational goals, and project a future
orientation. People living with HIV can detect important differences in the appropriateness of these
responses to disclosures, which should help communication scholars in developing recommendations for
future interventions.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 37
References
Agne, R., Thompson, T. L., & Cusella, L. R. (2000). Stigma in the line of face: Self disclosure of
patients’ HIV status to health care providers. Journal of Applied Communication Research, 28,
235-261.
Alonzo, A. A., & Reynolds, N. R. (1995). Stigma, HIV, and AIDS: An exploration and elaboration of a
stigma trajectory model. Social Science and Medicine, 41, 303-315.
Barbee, A. P., Derlega, V. J., Sherburne, S. P., & Grimshaw, A. (1998). Helpful and unhelpful forms of
social support for HIV-positive individuals. In V. J. Derlega & A. P. Barbee (Eds.), HIV and
social interaction (pp. 83-105). Thousand Oaks, CA: Sage.
Brashers, D. E., Neidig, J. L., Russell, J. A., Cardillo, L. W., Haas, S. M., Dobbs, L. K., et al. (2003). The
medical, personal, and social causes of uncertainty in HIV illness. Issues in Mental Health
Nursing, 24, 497-522.
Brashers, D. E., Hsieh, E., Neidig, J. L., & Reynolds, N. R. (2006). Managing uncertainty about illness:
Health care providers as credible authorities. In R. M. Dailey & B. A. Le Poire (Eds.), Applied
interpersonal communication matters: Family, health, and community relations (pp. 219-240).
New York: Peter Lang.
Brashers, D. E., Neidig, J. L., & Goldsmith, D. J. (2004). Social support and the management of
uncertainty for people living with HIV. Health Communication, 16, 305-331.
Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. Cambridge, England:
Cambridge University Press.
Brunig, J. L., & Kintz, B. L. (1997). Computational handbook of statistics (4th. ed.) New York:
Longman.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 38
Burleson, B. R. (1994). Comforting communication: Significance, approaches, and effects. In B. R.
Burleson, T. L. Albrecht, & I. G. Sarason (Eds.), Communication of social support: Messages,
interactions, relationships, and community (pp. 3-28). Thousand Oaks, CA: Sage.
Canary, D. J., & Spitzberg, B. H. (1987). Appropriateness and effectiveness in the perception of conflict
strategies. Human Communication Research, 13, 93-118.
Caughlin, J. P., Bute, J. J., Donovan-Kicken, E., Kosenko, K. A., Ramey, M. E., & Brashers, D. E. (in
press). Do message features influence reactions to HIV disclosures? A multiple goals perspective.
Health Communication.
Clark, R. A., & Delia, J. G. (1979). Topoi and rhetorical competence. Quarterly Journal of
Speech, 65, 187-206.
Chesney, M. A., & Smith, A. W. (1999). Critical delays in HIV testing and care: The potential role of
stigma. American Behavioral Scientist, 42, 1162-1174.
Coupland, N., Coupland, J., & Giles, H. (1991). Language, society, and the elderly: Discourse, identity,
and ageing. Oxford, UK: Blackwell.
Coupland, N., Henwood, K., Coupland, J., & Giles, H. (1989). Accommodating troubles-talk: The
young's management of elderly self-disclosure. In G. M. McGregor & R. White (Eds.),
Reception and response: Hearer creativity and the analysis of spoken and written texts (pp. 112-
144). London: Croom Helm.
Derlega, V. J., Lovejoy, D., & Winstead, B. A. (1998). Personal accounts of disclosing and concealing
HIV-positive test results: Weighing the benefits and risks. In V. J. Derlega & A. P. Barbee (Eds.),
HIV and social interaction (pp. 147-164). Thousand Oaks, CA: Sage.
Derlega, V. J., Winstead, B. A., & Folk-Barron, L. (2000). Reasons for and against disclosing HIV-
seropositive test results to an intimate partner: A functional perspective. In S. Petronio (Ed.),
Balancing the secrets of private disclosures (pp. 53–69). Mahwah, NJ: Erlbaum.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 39
Derlega, V. J., Winstead, B. A., Greene, K., Serovich, J., & Elwood, W. N. (2004). Reasons for HIV
disclosure/nondisclosure in close relationships: Testing a model of HIV-disclosure decision
making. Journal of Social and Clinical Psychology, 23, 747-767.
Dillard, J. P. (1990). A goal-driven model of interpersonal influence. In J. P. Dillard (Ed.), Seeking
compliance: The production of interpersonal influence messages (pp. 41-56). Scottsdale, AZ:
Gorsuch Scarisbrick.
Fleiss, J. L. (1981). Statistical methods for rates and proportions (2nd ed.). New York: Wiley.
Goldsmith, D. J. (2004). Communicating social support. New York: Cambridge University Press.
Goldsmith, D. J., & Fitch, K. (1997). The normative context of advice as social support. Human
Communication Research, 23, 454-476.
Goldsmith, D. J., McDermott, V. M., & Alexander, S. C. (2000). Helpful, supportive and sensitive:
Measuring the evaluation of enacted social support in personal relationships. Journal of Social
and Personal Relationships, 17, 369-391.
Gosling, S. D., Vazire, S., Srivastava, S. & John, O. P. (2004). Should we trust web-based studies?
American Psychologist, 59, 93-104.
Greene, K., Derlega, V. J., Yep, G. A., & Petronio, S. (2003). Privacy and disclosure of HIV in
interpersonal relationships. Mahwah, NJ: Erlbaum.
Greene, K., & Faulkner, S. L. (2002). Expected versus actual responses to disclosure in relationships of
HIV-positive African American adolescent females. Communication Studies, 53, 297-317.
Hatfield, J. D., & Weider-Hatfield, D. (1978). The comparative utility of three types of behavioral units
for interaction analysis. Communication Monographs, 45, 44-50.
Herek, G. M., Capitanio, J. P., & Widaman, K. F. (2002). HIV-related stigma and knowledge in the
United States: Prevalence and trends, 1991–1999. American Journal of Public Health, 92, 371-
377.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 40
Hullman, G. A. (2004). Interpersonal communication motives and message design logic: Exploring their
interaction on perceptions of competence. Communication Monographs, 71, 208-225.
Jackson, S. (1992). Message effects research: Principles of design and analysis. New York: Guilford
Press.
Jackson, S., & Brashers, D. E. (1994). Random factors in ANOVA. Thousand Oaks, CA: Sage.
Kunkel, A. D., Wilson, S. R., Olofowote, J., & Robson, S. (2003). Identity implications of influence
goals: Initiating, intensifying, and ending romantic relationships. Western Journal of
Communication, 67, 382-412.
Lambert, B. L., & Gillespie, J. L. (1994). Patient perceptions of pharmacy students' hypertension
compliance-gaining messages: Effects of message design logic and content themes. Health
Communication, 6, 311-325.
Le Poire, B. A. (1994). Attraction toward and nonverbal stigmatization of gay males and persons with
AIDS: Evidence of symbolic over instrumental attitudinal structures. Human Communication
Research, 21, 241-279.
MacGeorge, E. L., Gillihan, S. J., Samter, W., & Clark, R. A. (2003). Skill deficit or differential
motivation? Testing alternative explanations for gender differences in the provision of emotional
support. Communication Research, 29, 1-32.
O’Keefe, B. J. (1988). The logic of message design: Individual differences in reasoning about
communication. Communication Monographs, 55, 80-103.
O’Keefe, B. J. (1990). The logic of regulative communication: Understanding the rationality of message
designs. In J. P. Dillard (Ed.), Seeking compliance: The production of interpersonal influence
messages (pp. 87-104). Scottsdale, AZ: Gorsuch Scarisbrick.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 41
O’Keefe, B. J. (1991). Message design logic and the management of multiple goals. In K. Tracy (Ed.),
Understanding face-to-face interaction: Issues linking goals and discourse (pp. 131-150).
Hillsdale, NJ: Erlbaum.
O’Keefe, B. J. (1992). Developing and testing rational models of message design. Human
Communication Research, 18, 637-649.
O'Keefe, B. J., Lambert, B. L., & Lambert, C. A. (1997). Conflict and communication in a research and
development unit. In B. D. Sypher (Ed.), Case studies in organizational communication 2:
Perspectives on contemporary worklife (pp. 31-52). New York: Guilford Press.
O'Keefe, B. J., & McCornack, S. A. (1987). Message design logic and message goal structure: Effects on
perceptions of message quality in regulative communication situations. Human Communication
Research, 14, 68-92.
O'Keefe, B. J., & Shepherd, G. J. (1987). The pursuit of multiple objectives in face-to-face persuasive
interactions: Effects of construct differentiation on message organization. Communication
Monographs, 54, 396-419.
Peters, M. (2005, November). Message design logic and comforting communication in a chronic illness
context: Introducing a message elicitation task and adapted coding scheme. Paper presented at
the Annual Convention of the National Communication Association, Boston, MA.
Peterson, L. W., & Albrecht, T. L. (1996). Message design logic, social support, and mixed-status
relationships. Western Journal of Communication, 60, 291-309.
Rintamaki, L. R., Davis, T. C., Skripkauskas, S., Bennett, C. L., & Wolf, M. S. (2006). Social stigma
concerns and HIV medication adherence. AIDS Patient Care and STDs, 20, 359-368.
Saeki, M., & O’Keefe, B. J. (1994). Refusals and rejections: Designing messages to serve multiple goals.
Human Communication Research, 21, 67-102.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 42
Schrimshaw, E. W., & Siegel, K. (2002). HIV-infected mothers' disclosure to their uninfected children:
Rates, reason, and reactions. Journal of Social and Personal Relationships, 19, 19-43.
Serovich, J. M., Kimberly, J. A., & Greene, K. (1998). Perceived family member reaction to women's
disclosure of HIV-positive information. Family Relations, 47, 15-22.
Serovich, J. M., Mason, T. L., Bautista, D., & Toviessi, P. (2006). Gay men’s report of regret of HIV
disclosure to family, friends, and sex partners. AIDS Education and Prevention, 18, 132-138.
Sullivan, K. M. (2005). Male self-disclosure of HIV-positive serostatus to sex partners: A review of the
literature. Journal of the Association of Nurses in AIDS Care, 16, 33-47.
Vangelisti, A. L. (1993). Communication in the family: The influence of time, relational prototypes, and
irrationality. Communication Monographs, 60, 42-54.
Wilson, S. R. (2002). Seeking and resisting compliance: Why people say what they do when trying to
influence others. Thousand Oaks, CA: Sage.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 43
Table 1
Frequencies and Percentages of Messages with Different Design Logics by Sex of Participant
Sex ____________________________
Message Design Logic Male Female
Expressive 86 (36.60%) 84 (28.47%)
Conventional 118 (50.21%) 176 (59.66%)
Rhetorical 31 (13.19%) 35 (11.86%)
Note: Due to rounding, the percentages do not add to exactly 100.00 for the female participants.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 44
Table 2
Proportion of Responses of Various Design Logics for Each Condition
Design Logic
____________________________________________________ Condition Expressive Conventional Rhetorical Kernel .25ab .59 .16a Seeking Support .29 .61a .10 Avoid Negative Evaluation .32 .54 .15 Prevent Subsequent Disclosure .43ac .51 .06ab Maintain Relationship .28c .62b .11 Right to Know .40b .45ab .16b Note. Proportions in the same column that share subscripts differ significantly at p < .05. Due to rounding,
the proportions for some conditions do not add to 1.00.
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 45
APPENDIX
Coding Rules
General Rules Coding occurs at the level of the message, based on a holistic assessment of its overall characteristics. Some features will supercede others (e.g., if a message contains overt insults and hurtful comments, it must be an expressive message; if a message contains implicit or explicit blame, it cannot be a rhetorical message). Statements indicating what the participant would do should be coded as if the person did that act. For instance, “I would try to be supportive” counts as being supportive and “I would be there to listen” counts as an attempt to comfort by listening. Exclude what participants indicate that they would do at a later time (e.g., “From then on, I would pay attention to news about HIV treatment and encourage, support her through this period of time”). Exclude editorializing unless it appears that they are stating what they would say or do (e.g., do not consider “I would initially be shocked, because I could not imagine my brother with any serious illness;” but do consider “I would ask him if he was serious or if he was joking with me.”).
Expressive Design Logic General Principles Communication is a medium for expressing feelings
Clarity, openness, honesty, unimpeded expression Goal is self expression
Features that Guarantee Expressive Code
Noncontingent threats Insults and/or inappropriate/rude/hurtful comments Explicit blame (Implicit blame can still be conventional) Explicit criticism
Features that Expressive Messages Tend to Contain
Irrelevant content (irrelevant can be thought of in terms of the need to comfort the other) Marked redundancies Complaints that the hearer can do nothing about Inoffensive but inappropriate remarks Lack of response to explicit goal of discloser (i.e., not saying anything about the relationship when that is explicitly framed in the disclosure) Expression of the mental state of the speaker (use of emotional expression functions) Lies about mental state (emotional inhibition) Series of questions geared toward one’s own interests (i.e., not helpful/productive) Lack of coherence; disconnect between remarks (e.g., “stream of consciousness”) Invalidation of the discloser’s perspective Implication that the discloser is incompetent Unsolicited advice that is insulting or blaming Hints that the discloser is “toxic” or dangerous (e.g., remarks about contagion)
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x
Responses to HIV Disclosures 46
Conventional Design Logic
General Principles Communication is a game played cooperatively by social rules Appropriateness, control of resources, cooperativeness Given the normative goals in this context, a message must provide at least minimal comforting to qualify as conventional
Features that Conventional Messages Tend to Contain
Standard, acceptable, appropriate, or obligatory response(s) to the situation Typical content and structure Clearly identifiable core action that is being performed Implicit acknowledgement of the discloser’s goals Responses that relate directly to the situation at hand and connect to the discloser’s goals, without elaborating on them (e.g., attempt at comfort, denial of request for privacy) Offers of assistance or support without actually indicating a commitment to take on the situation together Expressive content that is more goal-oriented Use of social scripts or lines A message can still be conventional even with implicit blame as long as there are clear comforting attempts and nothing that makes it automatically expressive
Clarifications for Coding Advice that is intended to be helpful counts as assistance or support Statements indicating that the participant would hug the discloser count as support
Rhetorical Design Logic General Principles Communication is the creation and negotiation of social selves and situations
Flexibility, symbolic sophistication, depth of interpretation Goal is to negotiate social consensus
Features that Rhetorical Messages Tend to Contain
Definition of roles Creation of context Explicit definition of context Indication of a communal coping orientation (i.e., defining the situation as “our problem, our solution”) Legitimation of the discloser’s feelings or perspective Affirmation of the discloser’s goals Suggestions for accomplishing what the discloser wants Elaboration of how to achieve goals Attempts to achieve consensus (e.g., questions that seek the hearer’s agreement or opinion) Emphasis on careful listening Presence of identity, relational, or privacy messages in absence of an explicit framing of those goals Rational arguments (appraisal support) Persuasive attempts to convey one’s position
Features that Preclude Rhetorical Code
Implicit blame Dismissal of the discloser’s feelings Contingent threat
This is the accepted version of the following article: Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan-Kicken, E., & Bute, J. J. (2008). The Message Design Logics of Responses to HIV Disclosures. Human Communication Research, 34(4), 655–684, which was published in its final form at: http://dx.doi.org/10.1111/j.1468-2958.2008.00336.x